daskycodes / infer

Infer file and MIME type by checking the magic number signature

Infer

A dependency free library to infer file and MIME type by checking the magic number signature.

An elixir adaption of the infer rust library.

Installation

The package can be installed by adding infer to your list of dependencies in mix.exs:

def deps do
  [
    {:infer, "~> 0.2.6"}
  ]
end

The docs can be found at https://hexdocs.pm/infer.

Examples

Infer.get/1

Takes the binary file contents as argument and returns the Infer.Type.t() if the file matches one of the supported types. Returns nil otherwise.

iex> binary = File.read!("test/images/sample.png")
iex> Infer.get(binary)
%Infer.Type{extension: "png", matcher: &Infer.Image.png?/1, matcher_type: :image, mime_type: "image/png"}

Infer.get_from_path/1

Similar to Infer.get/1, but takes the file path as argument.

iex> Infer.get_from_path("test/images/sample.png")
%Infer.Type{extension: "png", matcher: &Infer.Image.png?/1, matcher_type: :image, mime_type: "image/png"}

Infer.is?/2

Takes the binary content and the file extension as arguments. Returns whether the file content is of the given extension.

iex> binary = File.read!("test/images/sample.png")
iex> Infer.is?(binary, "png")
true

Infer.mime?/2

Takes the binary content and the file extension as arguments. Returns whether the file content is of the given mime type.

iex> binary = File.read!("test/images/sample.png")
iex> Infer.mime?(binary, "image/png")
true

Infer.image?/1

Takes the binary file contents as argument and returns whether the file is an image or not.

iex> binary = File.read!("test/images/sample.png")
iex> Infer.image?(binary)
true

Infer.document?/1

Takes the binary file contents as argument and returns whether the file is a document (microsoft office, open office)

iex> binary = File.read!("test/docs/sample.xlsx")
iex> Infer.document?(binary)
true

Infer.Doc.docx?/1

Takes the binary file contents as arguments. Returns true if it's Microsoft Word Open XML Format Document (DOCX) data.

iex> binary = File.read!("test/docs/sample.docx")
iex> Infer.Doc.docx?(binary)
true

Supported Types

Image

MIME	Extension
image/jpeg	jpg
image/jp2	jp2
image/png	png
image/gif	image/web
image/x-canon-c32	cr2
image/tiff	tif
image/bmp	bmp
image/vnd.ms-photo	jxr
image/vnd.adobe.photoshop	psd
image/vnd.microsoft.icon	ico
image/heif	heif
image/avif	avif

Video

MIME	Extension
video/mp4	mp4
video/x-m4v	m4v
video/x-matroska	mkv
video/webm	webm
video/quicktime	mov
video/x-msvideo	avi
video/x-ms-wmv	wmv
video/mpeg	mpg
video/x-flv	flv

Audio

MIME	Extension
audio/midi	midi
audio/mpeg	mp3
audio/m4a	m4a
audio/ogg	ogg
audio/x-flac	flac
audio/x-wav	wav
audio/amr	amr
audio/aac	aac
audio/x-aiff	aiff

Document

MIME	Extension
application/msword	doc
application/vnd.openxmlformats-officedocument.wordprocessingml.document	docx
application/vnd.ms-excel	xls
application/vnd.openxmlformats-officedocument.spreadsheetml.sheet	xlsx
application/vnd.ms-powerpoint	ppt
application/vnd.openxmlformats-officedocument.presentationml.presentation	pptx
application/vnd.oasis.opendocument.text	odt
application/vnd.oasis.opendocument.spreadsheet	ods
application/vnd.oasis.opendocument.presentation	odp

Archive

MIME	Extension
application/epub+zip	epub
application/zip	zip
application/x-tar	tar
application/vnd.rar	rar
application/gzip	gz
application/x-bzip2	bz2
application/x-7z-compressed	7z
application/x-xz	xz
application/is_pdf	pdf
application/x-shockwave-flash	swf
application/rtf	rtf
application/octet-stream	eot
application/postscript	ps
application/vnd.sqlite3	sqlite
application/x-nintendo-nes-rom	nex
application/x-unix-archive	ar
application/x-compressed	Z
application/x-lzip	lz
application/x-rpm	rpm
application/dicom	dcm
application/zstd	zst
application/x-ole-storage	msi
application/x-google-chrome-extension	crx
application/vnd.ms-cab-compressed	cab
application/vnd.debian.binary-package	deb

Font

MIME	Extension
application/font-woff	woff
application/font-woff2	woff2
application/font-sfnt	ttf
application/font-sfnt	otf

Book

MIME	Extension
application/epub+zip	epub
application/x-mobipocket-ebook	mobi

Application

MIME	Extension
application/wasm	wasm
application/x-executable	elf
application/vnd.microsoft.portable-executable	exe
application/vnd.microsoft.portable-executable	dll
application/java	class
application/x-llvm	bc
application/x-mach-binary	mach
application/vnd.android.dex	dex
application/vnd.android.det	dey
application/x-x509-ca-cert	der
application/x-executable	obj

License

This project and is licensed under the MIT License - see the LICENSE file for details.

Copyright (c) 2021 Daniel Khaapamyaki

infer's Issues

Possible leaking file descriptor in `Infer.get_from_path/1,2`

Consider the implementation of Infer.get_from_path/1,2:

infer/lib/infer.ex

Lines 46 to 53 in 65beb86

	@spec get_from_path(binary()) :: Infer.Type.t() | nil
	def get_from_path(path, byte_size \\ 2048) do
	with {:ok, io_device} <- :file.open(path, [:read, :binary]),
	{:ok, binary} <- :file.read(io_device, byte_size) do
	:file.close(io_device)
	Enum.find(@matchers, & &1.matcher.(binary))
	end
	end

If :file.read/2 returns anything other than {:ok, data}, the file won't be closed. One possible solution is to use File.open/3 with a function as an argument because it ensures the file will be closed even in case of error:

def get_from_path(path, byte_size \\ 2048) do
  result = File.open(path, [:binary, :read], fn io_device ->
    case IO.binread(io_device, byte_size) do
      binary when is_binary(binary) -> Enum.find(@matchers, & &1.matcher.(binary)) 
      _other -> nil
    end
  end)

  case result do
    {:ok, %Infer.Type{} = type} -> type
    _other -> nil
  end
end

(this implementation would also fix #9)

`is_*` functions should have no prefix and end with a question mark

Note the convention in Elixir is to name functions/macros allowed in guards with the is_ prefix, such as is_list/1. If, however, the function/macro returns a boolean and is not allowed in guards, it should have no prefix and end with a question mark, such as Keyword.keyword?/1.

https://hexdocs.pm/elixir/Kernel.html#defguard/1

Infer functions fail to typecheck (functions don't exist)

Consider the following minimal example (or see the repo):

defmodule InferTypecheckIssue do
  def get(binary), do: Infer.get(binary)
  def get_from_path(path), do: Infer.get_from_path(path)
end

When running dialyzer (via dialyxir v1.4.2), it fails with errors:

$ mix dialyzer
[...snip...]
Total errors: 2, Skipped: 0, Unnecessary Skips: 0
done in 0m0.69s
lib/infer_typecheck_issue.ex:2:unknown_function
Function Infer.get/1 does not exist.
________________________________________________________________________________
lib/infer_typecheck_issue.ex:3:unknown_function
Function Infer.get_from_path/1 does not exist.
________________________________________________________________________________
done (warnings were emitted)
Halting VM with exit status 2

I must admit I have no idea why it happens. I thought that it has something to do with the infer's project config, so I looked at mix.exs but didn't find anything suspicious.

Invalid type spec for `Infer.get_from_path/1,2`

The type specification states that the function returns either Infer.Type.t() or nil, but it returns {:error, :file.posix()} if given a non-existing path:

iex(1)> Infer.get_from_path("/path/to/non/existing/file.jpeg")
{:error, :enoent}

Replace sample files with smaller filesize

Currently the sample files infer uses for testing are generic sample files.

To improve the speed of the test suite most of the files can be replaced by smaller files.

• Replace sample files with smaller files